1 /07/ $20.00 PHARMACOLOGICAL REVIEWS Vol. 59, No. 2 Copyright 2007 by The American Society for Pharmacology and Experimental Therapeutics 50441/ Pharmacol Rev 59: , 2007 Printed in U.S.A Pharmacological Treatment of the Overweight Patient GEORGE A. BRAY AND FRANK L. GREENWAY Pennington Biomedical Research Center, Louisiana State University System, Baton Rouge, Louisiana I. Introduction II. Using the currently available drugs III. Drugs approved by the U.S. Food and Drug Administration or the European Medicinal Evaluation Agency for treatment of overweight patients A. Drugs approved for long-term use Orlistat a. Mechanism of action b. Long-term studies c. Studies in special populations d. Meta-analysis of orlistat studies e. Safety considerations Sibutramine a. Mechanism of action b. Long-term studies c. Studies in special populations d. Meta-analysis of sibutramine studies e. Combining sibutramine and orlistat f. Dosage and safety considerations Rimonabant a. Mechanism of action b. Long-term studies c. Studies in special populations d. Safety considerations e. Other cannabinoid antagonists B. Drugs approved by the U.S. Food and Drug Administration for short-term treatment of overweight patients Sympathomimetic drugs a. Mechanism of action b. Clinical studies IV. Antidepressant and antiepileptic drugs that produce weight loss but are not approved by the U.S. Food and Drug Administration for weight loss A. Fluoxetine and Sertraline Mechanism of action Clinical studies B. Bupropion Mechanism of action Clinical studies for weight loss C. Topiramate Mechanism of action Clinical studies for weight loss Special situations D. Zonisamide Mechanism of action Clinical studies for weight loss Address correspondence to: Dr. George A. Bray, Pennington Center, 6400 Perkins Road, Baton Rouge, LA pbrc.edu This article is available online at doi: /pr
2 152 BRAY AND GREENWAY E. Lamotrigine Mechanism of action Clinical studies V. Drugs approved by the U.S. Food and Drug Administration for uses other than overweight A. Metformin Mechanism of action Clinical studies B. Pramlintide Mechanism of action Clinical studies C. Exenatide Mechanism of action Clinical studies with weight loss as a component D. Somatostatin Mechanism of action Clinical studies for weight loss E. Atomoxetine Mechanism of action Clinical study for weight loss F. Growth hormone and growth hormone fragment Mechanism of action Clinical studies on body composition VI. Drugs with clinical data or in clinical studies A. Leptin Mechanism of action Clinical studies B. Neuropeptide Y receptor antagonists Mechanism of action Clinical studies C. Serotonin 2C receptor agonists Mechanism of action Clinical studies D. Peptide YY Mechanism of action Clinical studies E. Oxyntomodulin Mechanism of action Clinical studies F. Pancreatic lipase inhibitor Mechanism of action Clinical studies G. Cholecystokinin H. Combination of bupropion and naltrexone (Contrave) I. Combination of bupropion and zonisamide (Empatic) J. Combination of topiramate and phentermine (Qnexa) VII. Drugs in the early phases of development A. Melanin-concentrating hormone receptor-1 antagonist B. Histamine-3 receptor antagonists C. Ghrelin antagonists D. Angiogenesis antagonists and fat cell antibodies VIII. Drugs and herbal medications no longer under investigation or withdrawn A. Ephedra B. 3 -Adrenergic agonists C. Bromocriptine D. Ecopipam E. Axokine IX. Over-the-counter medications
3 DRUG TREATMENT OF OBESITY 153 A. Orlistat B. Phenylpropanolamine X. Herbal products, functional foods, and nutriceuticals A. Interventions requiring special training Acupuncture/acupressure Homeopathy Hypnotherapy B. Minerals and metabolites Chromium picolinate Hydroxymethyl butyrate Pyruvate Conjugated linoleic acid Calcium C. Herbal dietary supplements Ephedra sinica Green tea extract Garcinia cambogia Yohimbine from Pausinystalia yohimbe Hoodia Citrus aurantium (bitter orange) Ayurvedic preparations D. Fibers Chitosan Glucomannan Guar gum Plantago psyllium XI. Conclusions References I. Introduction To use medications properly for treatment of the overweight patient, it is important to start with a framework based on the realities associated with this treatment. These are briefly summarized below: Overweight is due to an imbalance between energy intake and energy expenditure. Drugs can either reduce food intake or increase energy expenditure. Drug treatment does not cure the overweight patient. The therapeutic armamentarium of physicians is limited to only a few drugs. The use of drugs labors under the negative halo of treatment mishaps. Drugs do not work when they are not taken; when 1 Abbreviations: DPP, Diabetes Prevention Program; BMI, body mass index; GLP-1, glucagon-like peptide-1; CI, confidence interval; HDL, high-density lipoprotein; LDL, low-density lipoprotein; GI, gastrointestinal; NPY, neuropeptide Y; POMC, proopiomelanocortin; VLCD, very low-calorie diet; FDA, U.S. Food and Drug Administration; PYY, peptide YY; 5-HT, 5-hydroxytryptamine (serotonin); TNP- 470, O-(chloroacetyl-carbamoyl); L , (R)-N-[4-[2-[[2-hydroxy- 2-(3-pyridinyl)ethyl]amino]ethyl]-phenyl]-4-[4-[4-(trifluoromethyl)- phenyl]thiazol-2-yl]-benzenesulfonamide, dihydrochloride. drugs are stopped weight regain is the expected outcome. Weight loss plateaus during continued treatment when compensatory mechanisms come into play to counterbalance the effect of the drug. Monotherapy usually produces weight loss in the range of 10% (5% better than placebo). Frustration with the failure to continue to lose weight often leads to discontinuation of therapy and then to weight regain with labeling of the drug as a failure. Physicians have several strategies for confronting the problems of the overweight patient. The physician can counsel the patient that he or she is concerned about the patient s current level of body weight and can initiate treatment if the patient is interested. Alternatively, if a physician feels uncomfortable with addressing overweight in patients, he or she can ignore the problem and hope that the patient will not raise the issue. Or, finally the physician can wait until the complications of excess weight manifest themselves as diabetes, dyslipidemia, hypertension, or other disorders and then institute appropriate therapy for each of these medical problems. With the current high-quality therapies available to treat diabetes mellitus, dyslipidemia, and hypertension, many physicians would prefer this latter strategy.
4 154 BRAY AND GREENWAY However, if medical treatment of the overweight patient were more effective, physicians might prefer to treat the excess weight and thus delay the onset of the problems related to overweight. This strategy was the basis for the long-term Diabetes Prevention Program (DPP) 1 and the Swedish Obese Subjects Study. In the DPP, the onset of new cases of diabetes among individuals with impaired glucose tolerance was reduced 55% during an average follow-up of 3.2 years in the group who lost weight compared with the control group who did not lose weight (Knowler et al., 2002). In the Swedish Obese Subjects Study, the incidence of new cases of diabetes was reduced to zero over 2 years in patients who lost weight and maintained a weight loss of 12%, compared with an incidence of 8.5% for new cases of diabetes in those who did not lose weight (Sjostrom et al., 2004). Thus, effective treatment of the overweight patient at risk for diabetes or hypertension can reduce the risk of developing these serious diseases in the future. One reason most physicians are reluctant to treat overweight patients is that the treatments are limited in number and effectiveness. At the time this article was written, there were only two drugs approved by the U.S. Food and Drug Administration for long-term use. As monotherapy, these agents can produce an overall weight loss of 8 to 10% among patients who continue to take the medication for 6 months. However, to achieve the elimination of new cases of diabetes as noted above, the weight loss needs to exceed 12%, a goal that is not usually achieved with current monotherapy. Thus, there is a great need for new drugs to be used as monotherapy and probably in combinations when prevention fails. Both physicians and patients know that overweight is a stigmatized disease (Puhl and Brownell, 2003). One commonly held view is that overweight people are lazy and weak-willed. If fat people just had willpower, they would push themselves away from the table and not be overweight. This widely held view is shared by the public and by health professionals alike. The clamoring of women to be lean and well proportioned supports this view. The declining relative weight of centerfold models in Playboy magazine and of women who are winners of the Miss America contest in the latter part of the 20th century also supports this view. Many physicians just do not like to see overweight patients come into their offices. This attitude poses a major challenge to any efforts to improve the lot of people who are overweight. There can be both medical and cosmetic (self-image) benefits to weight loss. However, they do not necessarily occur together. For example, a 10% weight loss, which would be clinically significant for a 300-pound (145-kg) person, would only reduce body weight by 30 to 270 pounds, a weight change that many people might not notice and would not be considered a cosmetic success. At the other extreme, a 10% weight loss for an individual weighing 150 pounds would lower his or her weight to 135 pounds, which would have a very positive impact on self-image. We also know that cosmetically significant weight losses may not produce clinically significant effects. After liposuction that removed 7% of body weight, there were no improvements in health-related risk factors. These distinctions are shown in Table 1. Three other issues aggravate the problem of treating overweight patients. The first is the negative halo that surrounds the use of appetite suppressants because amphetamine is addictive. There was never any evidence that dexfenfluramine was addictive. Nonetheless, the drug was scheduled by the U.S. Drug Enforcement Agency as a Schedule IV drug because, on paper, it had chemical similarities to amphetamine. The second issue is the concern about the plateau of body weight that is reached when homeostatic mechanisms in the body come into play and stop further weight loss. There is an analogy with treatment of hypertension. When an antihypertensive drug is given, blood pressure drops and then stops falling within a few weeks to reach a plateau at a new lower level. The antihypertensive drug has not lost its effect when the plateau occurs, but its effect is being counteracted by physiological mechanisms designed to maintain blood pressure. In the treatment of overweight patients, a similar plateau in body weight is often viewed as a therapeutic failure for the weight loss drug. This is particularly so when weight is regained after the drug is stopped. These attitudes and biases need to change before any effective new therapy will become widely accepted. The final issue is the toxicity associated with many antiobesity drugs. The disaster that occurred for some of the patients who took the combination of fenfluramine and phentermine is one example (others are listed in Table 2). Aortic regurgitation occurred in up to 25% of the patients treated with this combination of drugs and led many physicians to say, I told you so and I m certainly glad I didn t use those drugs. This issue has largely subsided with time, but there will always remain a residue of concern among some physicians and among regulators about the potential problems that might sur- TABLE 1 Cosmetically significant versus clinically significant weight loss Type of Procedure Weight Loss Clinically Significant Cosmetically Significant Diet/exercise 10%; from 300 to 270 lb Yes No 10%; from 200 to 180 lb Yes Probably not 10%; from 150 to 135 lb Yes Yes Liposuction 7%; from 220 to 200 lb No Probably not 7%; from 160 to 149 lb No Yes Surgery (gastric bypass) 40%; from 264 to 165 lb Yes Yes
5 DRUG TREATMENT OF OBESITY 155 TABLE 2 Unintended consequences of some treatments for obesity Year Drug Consequence 1893 Thyroid Hyperthyroidism 1932 Dinitrophenol Cataracts/neuropathy 1937 Amphetamine Addiction 1968 Rainbow pills Deaths: arrhythmias 1985 Gelatin diets Cardiac deaths 1997 Phentermine/fenfluramine Valvulopathy 1998 Phenylpropanolamine Strokes 2003 Ma huang Heart attacks/strokes face when new treatments for overweight are made available to the public. Although the drug treatment of overweight patients has at least a century-long history (Colman, 2005), progress in drug discovery was given a new impetus by the discovery of leptin in 1994 (Zhang et al., 1994). This peptide demonstrated that overweight can be caused by a hormone deficiency and be reversed by replacement of that hormone (Halaas et al., 1995; Maffei et al., 1995; Farooqi et al., 2002). Even before the discovery of leptin, overweight had been declared to be a chronic disease by a National Institutes of Health Consensus Conference in 1985 (Bray, 2004). In the 20th century, bad eating habits were considered a primary cause for overweight. Because some bad habits can be behaviorally extinguished over a 12-week period of time, overweight medications approved before 1985 were approved for periods up to 12 weeks as an adjunct to a lifestyle change program. Equating overweight with bad habits and the stigmatization of obesity slowed the use of overweight medications chronically, as is done with medications for other chronic diseases (Puhl and Brownell, 2003). With the recognition that longer-term therapy was needed, clinical trials have been extended in length. Since 1990, only three medications have been approved for the chronic treatment of overweight, and one of them, dexfenfluramine, was withdrawn 2 years later (Anonymous, 1996). II. Using the Currently Available Drugs In this article, we review the field of drug therapy for the overweight patient. Table 3 lists the drugs that are available and whether they are approved for long-term use by the U.S. Food and Drug Administration or are restricted (scheduled) by the U.S. Drug Enforcement Agency on the basis of the belief that there is risk of abuse from the drug. For individuals desiring more detail or additional guidance in the use of medications to treat overweight, information can be found in a variety of sources (Bray and Greenway, 1999; National Heart, Lung, and Blood Institute and North American Association for the Study of Obesity, 2000; Haddock et al., 2002; Yanovski and Yanovski, 2002; Kim et al., 2003; Padwal et al., 2004, 2005; Colman, 2005; Li et al., 2005; Snow et al., 2005; Vettor et al., 2005). As a guide for the use of medications, we will use an algorithm that was described by the National Heart, Lung, and Blood Institute. The first step in this algorithm is to measure height, weight, and waist circumference to establish the body mass index (BMI) and the degree of central adiposity. If the BMI, ([weight in kilograms divided by the square of the height in meters] or [weight in pounds divided by square of the height (inches)] 703] is 30 kg/m 2 the patient is by definition obese and can be considered for medications. Overweight individuals with a BMI between 27 and 30 kg/m 2 may also be considered if they have diabetes, hypertension, sleep apnea, or another medical condition that would benefit from weight loss. Waist circumference is also an important indicator of risk from excess fat. The currently recommended upper limit for waist circumference in the United States is 102 cm (40 inches) for a man and 88 cm (35 inches) for a woman. A recent proposal from the International Diabetes Federation requires the presence of central adiposity to diagnose the metabolic syndrome and uses values for waist circumference 80 cm for females and 94 cm for males. Another important initial step in evaluating the overweight patient is to assess associated (comorbid) conditions by measuring blood pressure, glucose, lipids and, when indicated, by performing other tests. With results from this laboratory panel and waist circumference, the metabolic syndrome can be diagnosed. This is best done with the criteria from the National Cholesterol Education Panel Adult Treatment III Guidelines (Table 4). Once the patient has been established as an appropriate candidate to lose weight and he or she is motivated to do so, the next step is to set a weight loss goal. Most patients have an unrealistic view of how much weight they can lose. For them, a weight loss of 15% is often viewed as a failure. In contrast, weight loss using monotherapy with the drugs that are currently available is not usually 10%. It is, thus, important for physician and patient alike to set a weight loss goal for initial therapy that is not 10% and to set a lower limit for weight loss of 5%, which will suggest that an alternative strategy is needed. The next step is to be certain that the patient is ready to lose weight. With use of ideas from psychology, the patient must be ready to work on weight loss as opposed to not yet thinking about the problem. Once the weight goal is established and patients are prepared to take charge of their weight loss program, the next steps are to help them develop lifestyle changes that will benefit their program. The most important of these is monitoring what they eat, where they eat it, and under what circumstances they eat. A second element is to provide advice on diet and physical activity. Replacing voluntary choices with portion-controlled choices at one or more meals can be helpful. There are frozen foods, ready-to-make food items, and canned meal replacements that can be used for this purpose. Patients also need more exercise. One strategy is to have them get a step-counter and record the number of steps they take with the goal of
6 156 BRAY AND GREENWAY TABLE 3 Drugs producing weight loss that are approved by the U.S. Food and Drug Administration Drug Name Trade Name(s) Dosage U.S. Drug Enforcement Agency Schedule Side Effects and Comments Cannabinoid receptor antagonist Rimonabant Accomplia 20 mg/day N.A. Depressive symptoms, nausea, diarrhea; approved by the Center for Proprietary Medicinal Products in Europe approval by FDA pending Pancreatic lipase inhibitor Orlistat Xenical 120 mg t.i.d. before meals Norepinephrine-serotonin reuptake inhibitor Sibutramine Sympathomimetic drugs Diethylpropion Meridia Reductil Tenuate Tepanil Tenuate Dospan None Daily vitamin pill in the evening; may interact with cyclosporine 5 15 mg/day IV Raises blood pressure slightly; do not use with monoamine oxidase inhibitors, selective serotonin reuptake inhibitors, sumatriptan, dihydroergotamine, meperidine, methadone, pentazocine, fentanyl, lithium, or tryptophan 25 mg t.i.d. 25 mg t.i.d. 75 mg in A.M. Phentermine Standard release: mg t.i.d. IV Adipex-P Fastin Obenix Oby-Cap Oby-Trim Zantryl Benzphetamine Slow release: Ionamin Didrex mg/day in A.M. IV All sympathomimetic drugs are similar; do not use with monoamine oxidase inhibitors, guanethidine, alcohol, sibutramine, or tricyclic antidepressants mg 1 3 times/day III All sympathomimetic drugs are similar; do not use with monoamine oxidase inhibitors, guanethidine, alcohol, sibutramine, or tricyclic antidepressants Phendimetrazine Standard release: Bontril PDM Plegine X-Trozine Slow release: Bontril Prelu-2 X-Trozine 35 mg t.i.d. before meals III All sympathomimetic drugs are similar; do not use with monoamine oxidase inhibitors, guanethidine, alcohol, sibutramine, or tricyclic antidepressants 105 mg/day in A.M. TABLE 4 The National Cholesterol Education Program Adult Treatment Panel III and the International Diabetes Federation criteria for diagnosis of the metabolic syndrome Criterion ATP-III Modified Criteria International Diabetes Federation Waist circumference Female 35 inches 88 cm 31 inches 80 cm Male 40 inches 102 cm 37 inches 94 cm HDL cholesterol Female 50 mg/dl 1.29 mmol/l 50 mg/dl 1.29 mmol/l Male 40 mg/dl 1.03 mmol/l 40 mg/dl 1.03 mmol/l Glucose 110 mg/dl 6.2 mmol/l 100 mg/dl 5.6 mmol/l Blood pressure 130/85 mm Hg 130/85 mm Hg gradually increasing this number to 10,000 steps/day. In a review of lifestyle treatment used with pharmacotherapy in randomized clinical trials Poston et al. (2001) found that balanced-deficit diets were used in 40.7%, low-calorie diets in 25% and self-monitoring behavioral strategies in 23.1% of patients (Poston et al., 2001). When a patient returns to you, establish whether he or she has met the goals. If so, the patient continues as is, but if after 3 months he or she fails to meet the goals, then medications may be considered. The next step is to discuss the pros and cons of medication with the patient. An algorithm from the American College of Physicians (Snow et al., 2005) recommends six medications: orlistat, sibutramine,
7 phentermine, diethylpropion, fluoxetine, and bupropion. The first four have been approved by the U.S. Food and Drug Administration for treatment of overweight patients, but fluoxetine and bupropion have not, and they should not be used primarily for this purpose. In our view, fluoxetine and bupropion should only be used for weight loss in special situations. Fluoxetine is appropriate for the overweight patient who is depressed. Bupropion may also be helpful in reducing or preventing weight gain when people try to stop smoking and when they are depressed. DRUG TREATMENT OF OBESITY 157 III. Drugs Approved by the U.S. Food and Drug Administration or the European Medicinal Evaluation Agency for Treatment of Overweight Patients A. Drugs Approved for Long-Term Use 1. Orlistat. a. Mechanism of action. Orlistat is a lipase inhibitor. In pharmacological studies, it was shown to be a potent selective inhibitor of pancreatic lipase and to thus reduce the intestinal digestion of fat. The drug has a dose-dependent effect on fecal fat loss, increasing it to 30%. Thus, orlistat is recommended to be used with a diet that has 30% of its energy as fat. Orlistat has little effect in subjects eating a low-fat diet, as might be anticipated from its mechanism of action. In single-dose randomized and placebo-controlled studies, 120 mg of orlistat was shown to increase glucagon-like peptide-1 (GLP-1) and C-peptide more than placebo (Damci et al., 2004), to increase fecal fat loss but decrease the acute increase in cholecystokinin (O Donovan et al., 2003), but not to influence the behavioral measures of satiety (Goedecke et al., 2003). b. Long-term studies. Results of a number of 1- to 2-year long-term clinical trials with orlistat have been published. The results of a 2-year trial are shown in Fig. 1 (Sjostrom et al., 1998). The trial consisted of two parts. In the 1st year, patients received a hypocaloric diet calculated to be 500 kcal/day less than the patient s requirements. During the 2nd year, the diet was calculated to maintain body weight. By the end of year 1, the placebo-treated patients lost 6.1% of their initial body weight and the drug-treated patients lost 10.2%. The patients were randomized again at the end of year 1. Those switched from orlistat to placebo gained weight from 10 to 6% below baseline. Those switched from placebo to orlistat lost weight from 6 to 8.1% below baseline, which was essentially identical to the 7.9% weight loss in the patients treated with orlistat for the full 2 years. In a second 2-year study, 892 patients were randomized to orlistat or placebo (Davidson et al., 1999). One group received placebo throughout the 2 years (97 patients), and a second group received orlistat (120 mg three times per day) for 2 years (109 patients). At the end of 1 year, the dose for two-thirds of the group treated with orlistat for 1 FIG. 1. Double-blind, randomized clinical trial of orlistat versus placebo with a rerandomization of participants after the 1st year. Reprinted from The Lancet, vol. 352, Sjostrom et al., Randomized Placebo-Controlled Trial of Orlistat for Weight Loss and Prevention of Weight Regain in Obese Patients: European Multicenter Orlistat Study Group, pp , copyright 1998, with permission from Elsevier. year was changed to 60 mg three times per day (102 patients), and the others were switched to placebo (95 patients) (Davidson et al., 1999). After 1 year, the weight loss was 8.67 kg in the orlistat-treated group and 5.81 kg in the placebo group (P 0.001). During the 2nd year, those switched to placebo after 1 year reached the same weight as those treated with placebo for 2 years ( 4.5% in those treated with placebo for 2 years and 4.2% in those switched to placebo during year 2). In a third 2-year study, 783 patients remained in the placebo or orlistat-treated groups at 60 or 120 mg three times per day for the entire 2 years (Rossner et al., 2000). After 1 year with a weight-loss diet, the placebo group lost 7 kg, which was significantly less than the 9.6 kg lost by the group treated with orlistat 60 mg three times daily or the 9.8 kg lost by the group treated with 120 mg of orlistat three times daily. During the 2nd year, when the diet was liberalized to a weight maintenance diet, all three groups regained some weight. At the end of 2 years, the patients in the placebo group were 4.3 kg below baseline, the patients treated with 60 mg of orlistat three times per day were 6.8 kg below baseline, and the patients who took 120 mg of orlistat three times per day were 7.6 kg below baseline. The final 2-year trial evaluated 796 subjects in a general-practice setting (Hauptman, 2000). After 1 year of treatment with 120 mg of orlistat three times per day, the orlistat-treated patients (n 117) lost 8.8 kg, compared with 4.3 kg in the placebo group (n 91). During the 2nd year, when the diet was liberalized to maintain body weight, both groups regained some weight. At the end of 2 years, the orlistat group was 5.2 kg below their baseline weight
8 158 BRAY AND GREENWAY compared with 1.5 kg below baseline for the group treated with placebo. The pooled 2-year data from these four studies are shown in Fig. 2. This figure contains information on both the 120- and 60-mg-three-times-a-day doses. It is clear that there is a dose response. The maximal weight loss was achieved between 6 and 9 months, and then there was a slow regain in all of the groups during the rest of the study. The results of a 4-year double-blind, randomized, placebo-controlled trial with orlistat have also been reported (Torgerson et al., 2004). A total of 3304 overweight patients, 21% of whom had impaired glucose tolerance, were included in this Swedish trial (Fig. 3). The lowest body weight was achieved during the 1st year: 11% below baseline in the orlistat-treated group and 6% below baseline in the placebo-treated group. Over the remaining 3 years of the trial, there was a small regain in weight, such that by the end of 4 years, the orlistat-treated patients were 6.9% below baseline compared with 4.1% for those receiving placebo. The trial also showed a 37% reduction in the conversion of patients from impaired glucose tolerance to diabetes; essentially all of this benefit occurred in the patients with impaired glucose tolerance when they were enrolled into the trial. Weight maintenance with orlistat was evaluated in a 1-year study (Hill et al., 1999). Patients were enrolled if they had lost 8% of their body weight over 6 months while eating a 1000 kcal/day (4180 kj/day) diet. The 729 patients were randomized to receive placebo or orlistat at 30, 60, or 120 mg three times per day for 12 months. At the end of this time, the placebotreated patients regained 56% of their body weight, compared with 32.4% in the group treated with 120 mg of orlistat three times per day. The other two doses of orlistat were no different from placebo in preventing the regain of weight. c. Studies in special populations. i. Diabetic patients. Patients with diabetes treated with orlistat, 120 mg three times daily for 1 year, lost FIG. 2. Two-year pooled data comparing orlistat and 120 and 60 mg three times a day and placebo. From Hoffmann-LaRoche data, with permission from Dr. Jonathan Hauptman. FIG. 3. Effect of orlistat on body weight in a 4-year randomized, placebo-controlled clinical trial. more weight than the placebo-treated group (Hollander et al., 1998; Kelley et al., 2002; Miles et al., 2002). The subjects with diabetes also showed a significantly greater decrease in hemoglobin A 1c levels. In another study of orlistat and weight loss, investigators pooled data on 675 subjects from three of the 2-year studies described previously in which glucose tolerance tests were available (Heymsfield et al., 2000). During treatment, 6.6% of the patients taking orlistat converted from a normal to an impaired glucose tolerance test, compared with 10.8% in the placebo-treated group. None of the orlistat-treated patients who originally had normal glucose tolerance developed diabetes, compared with 1.2% in the placebo-treated group. Of those who initially had normal glucose tolerance, 7.6% in the placebo group but only 3% in the orlistat-treated group developed diabetes. The effect of orlistat in preventing diabetes has been assessed in a 4-year study (Torgerson et al., 2004). In this trial weight was reduced by 2.8 kg (95% CI kg) compared with placebo, and the conversion rate to diabetes was reduced from 9 to 6% for a relative risk reduction of 0.63 (95% CI ) (Padwal et al., 2005). ii. Metabolic syndrome and lipids. In a further analysis, patients who participated in the studies described above were divided into the highest and lowest quintiles for triglycerides and HDL cholesterol levels (Reaven et al., 2001). Those with high triglyceride and low HDL cholesterol levels were labeled syndrome X, and those with the lowest triglyceride levels and highest HDL cholesterol levels were the nonsyndrome X controls. With this classification, there were almost no men in the nonsyndrome X group, compared with an equal sex breakdown in the syndrome X group. In addition, the syndrome X group had slightly higher systolic and diastolic blood pressure levels and a nearly 2-fold higher level of fasting insulin. Besides weight loss, the only difference between the placebo and orlistat-treated patients was the decrease in LDL cholesterol levels in the
9 DRUG TREATMENT OF OBESITY 159 patients treated with orlistat. However, the syndrome X subgroup showed a significantly greater decrease in triglyceride and insulin levels than those without syndrome X. Levels of HDL cholesterol increased more in the syndrome X group, but LDL cholesterol levels showed a smaller decrease than that in the nonsyndrome X group. All of the clinical studies with orlistat have shown significant decreases in serum cholesterol and LDL cholesterol levels that usually are greater than decreases that can be accounted for by weight loss alone (Bray and Greenway, 1999). One study showed that orlistat reduces the absorption of cholesterol from the GI tract, thus providing a mechanism for the clinical observations (Mittendorfer et al., 2001). iii. Studies in children. A multicenter trial tested the effect of orlistat in 539 obese adolescents (Chanoine et al., 2005). Subjects were randomized to placebo or 120 mg of orlistat three times a day and a mildly hypocaloric diet containing 30% fat. By the end of the study BMI decreased 0.55 kg/m 2 in the drug-treated group but increased 0.31 kg/m 2 in the placebo group. By the end of the study, weight increased by only 0.51 kg in the orlistat-treated group, compared with 3.14 kg in the placebo-treated group (Fig. 4). This difference was due to differences in body fat. The side effects were gastrointestinal in origin, as expected from the mode of action of orlistat. A second small 6-month randomized clinical trial from a single site failed to show a difference resulting from treatment with orlistat in a population of 40 adolescents (Maahs et al., 2006). d. Meta-analysis of orlistat studies. Several metaanalyses of orlistat have been published (Haddock et al., 2002; Avenell et al., 2004; Li et al., 2005). By pooling six studies Haddock et al. (2002) estimated the weight loss in patients treated with orlistat to be 7.1 kg (range 4.0 to 10.3 kg) compared with 5.02 kg (range 3.0 to 6.1 kg) for the placebo-treated groups. In the metaanalysis of Li et al. (2005), the overall mean difference between orlistat and placebo after 12 months of therapy in 22 studies was 2.70 kg (95% CI 3.79 to 1.61 kg). Because this analysis included diabetic and nondiabetic subjects, we have summarized the data from the five 2-year studies in Table 5. In another meta-analysis of orlistat, 8-year-long studies, only one of which was in diabetic subjects, examined the effects of weight loss at 1 and 2 years and on the various laboratory and clinical responses. The overall effect of orlistat on weight loss at 12 months using the weighted mean difference was 3.01 kg (95% CI 3.48 to 2.54 kg) (Table 6). After 24 months, the overall effect of orlistat on weight loss was 3.26 kg (95% CI 4.15 to 2.37 kg). In terms of weight maintenance, the overall effect of orlistat after 12 months was 0.85 kg (95% CI 1.50 to 0.19 kg) (Davidson et al., 1999; Hill et al., 1999; Hauptman, 2000; Rossner et al., 2000). The pooled data show significant overall effects after 1 year of treatment on the change in cholesterol [ 0.34 mm (95% CI 0.41 to 0.027)] (n 7 studies), the change in LDL cholesterol [ 0.29 mm (95% CI 0.34 to 0.24)] (n 7 studies), the change in HDL cholesterol [ 0.03 mm (95% CI 0.05 to 0.01)] (n 6 studies), the change in triglycerides [0.03 mm (95% CI 0.04 to 0.10)] (n 6 studies), the change in hemoglobin A1c [ 0.17% (95% CI 0.24 to 0.10] (n 3 studies) (Hollander et al., 1998; Lindgarde, 2000; Broom et al., 2002), the change in systolic blood pressure [ 2.02 mm Hg (95% CI 2.87 to 1.17)] (n 7 studies), and the change in diastolic blood pressure [ 1.64 mm Hg (95% CI 2.20 to 1.09)] (n 7 studies)]. In a meta-analysis focused on the use of orlistat in diabetics Norris et al. (2004) reported a weighted mean difference in favor of orlistat of 2.6 kg (95% CI 3.2 to 2.1) after 52 to 57 weeks of treatment. e. Safety considerations. Orlistat is not absorbed to any significant degree from the gastrointestinal tract, and its side effects are thus related to the blockade of triglyceride digestion in the intestine (Zhi et al., 1999). Fecal fat loss and related GI symptoms are common initially, but they subside as patients learn to use the drug (Bray and Greenway, 1999). The quality of life in patients treated with orlistat may improve despite concerns about GI symptoms. Orlistat can cause small but significant decreases in fat-soluble vitamins. Levels usually remain within the normal range, but a few patients may need vitamin supplementation. Because it is impossible to tell which patients need vitamins, it is wise to TABLE 5 Meta-analysis of studies with long-term use of orlistat Adapted from Li et al. (2005). FIG. 4. Effect of orlistat on body mass index in a randomized, placebocontrolled clinical trial of orlistat in adolescents. Reproduced from The Journal of the American Medical Association, volume 293, pp Copyright 2005 American Medical Association. Reference Davidson et al. (1999) Hauptman (2000) Rossner et al. (2000) Sjostrom et al. (1998) Torgerson et al. (2004) Mean 95% CI 2.95 ( 4.45 to 1.45) 3.80 ( 5.37 to 2.23) 3.00 ( 4.17 to 1.83) 4.20 ( 5.26 to 3.14) 4.17 ( 4.60 to 3.74)
10 160 BRAY AND GREENWAY Adapted from Avenell et al. (2004). TABLE 6 Meta-analysis of the studies using orlistat Reference Treatment Placebo n Mean S.D. n Mean S.D. Weight Weighted Mean Difference % 95% CI Sjostrom et al. (1998) ( 5.35 to 3.05) Hollander et al. (1998) ( 3.54 to 1.29) Davidson et al. (1999) ( 4.45 to 1.45) Rossner et al. (2000) ( 4.30 to 1.50) Hauptman (2000) ( 5.31 to 2.67) Lindegarde (2000) ( 2.69 to 0.09) Finer et al. (2000) ( 3.73 to 0.23) Broom et al. (2002) ( 4.79 to 2.21) provide a multivitamin routinely with instructions to take it before bedtime. Orlistat does not seem to affect the absorption of other drugs, except cyclosporin. 2. Sibutramine. a. Mechanism of action. Sibutramine is a highly selective inhibitor for the reuptake at nerve endings of norepinephrine and serotonin and, to a lesser degree, dopamine. In preclinical experimental and clinical studies, it reduced food intake. In a double-blind placebocontrolled 2-week trial, a 30-mg/day dose of sibutramine reduced food intake by 23% on day 7 and 26% on day 14 relative to placebo and also decreased the percentage of fat eaten. A smaller dose of 10 mg also significantly reduced food intake at 14 days (Rolls et al., 1998). The effect of sibutramine on food intake has also been examined over a longer period of time (Barkeling et al., 2003). The first 2 weeks of this 10-month trial were conducted in a double-blind, randomized, placebo-controlled, crossover design. Participants then entered a 10-month openlabel trial with repeat food intake at the end. There was a 16% reduction in energy intake at the test lunch in the first part of the study (after 2 weeks). Ten months later, there was still a 27% reduction compared with participants preweight loss placebo-treatment food intake. In animals, sibutramine also stimulates thermogenesis, but there are conflicting data in humans (Hansen et al., 1998; Seagle et al., 1998). Mechanistic studies have shown that the effect of sibutramine can be mimicked by combining a selective serotonin reuptake inhibitor (fluoxetine) with a selective norepinephrine reuptake inhibitor (nisoxetine). When injected alone these specific reuptake inhibitors do not replicate the reduction of food intake produced by sibutramine (Jackson et al., 1997). Sibutramine treatment of experimental animals increased the activity of the sympathetic nervous system and attenuated the rise in NPY and fall in POMC in the arcuate nucleus in energy-restricted rats indicating that this drug influences both monoamine and peptidergic pathways involved in food intake (Levin and Dunn- Meynell, 2000). b. Long-term studies. Sibutramine has been approved by the U.S. Food and Drug Administration for long-term use in the treatment of overweight patients. Sibutramine has been evaluated extensively in several multicenter trials lasting 6 to 24 months. In a 6-month dose-ranging study of 1047 patients, 67% treated with sibutramine achieved a 5% weight loss from baseline, and 35% lost 10% (Bray and Greenway, 1999). There was a clear dose-response effect in this 24-week trial, and patients regained weight when the drug was stopped, indicating that the drug remained effective when used. Data from this multicenter trial are shown in Fig. 5 (Bray et al., 1999). In a 1-year trial of 456 patients who received sibutramine (10 or 15 mg/day) or placebo, 56% of those who stayed in the trial for 12 months lost at least 5% of their initial body weight, and 30% of the patients lost 10% of their initial body weight while taking the 10-mg dose (Smith and Goulder, 2001). Three trials have assessed the value of using sibutramine to prevent regain of body weight (Apfelbaum et al., 1999; James et al., 2000; Mathus-Vliegen, 2005). In a multicenter trial, participants were initially given a very low-calorie diet (VLCD) for 6 weeks to induce weight loss (Apfelbaum et al., 1999). Of the initial 181 subjects enrolled, 142 were randomized to either 10 mg/day of sibutramine or placebo after losing 6 kg or more with the VLCD. After another 12 months, those receiving the FIG. 5. Six-month randomized, placebo-controlled dose-ranging trial with sibutramine at six doses and placebo. Reproduced from Bray et al. (1999) with permission.
11 DRUG TREATMENT OF OBESITY 161 drug lost an additional 6.4 kg compared with a small weight gain of 0.2 kg for those receiving placebo. The authors concluded that sibutramine had effectively enhanced the initial weight loss and maintained it for an additional 12 months. This was followed by the Sibutramine Trial of Obesity Reduction and Maintenance (STORM) Trial that lasted 2 years and provided evidence for weight maintenance (James et al., 2000) (Fig. 6). Seven centers participated in this study, in which patients were initially enrolled in an open-label phase and treated with 10 mg/day of sibutramine for 6 months. Of the patients who lost 8 kg, two-thirds were then randomized to sibutramine and one-third to placebo. During the 18-month double-blind phase of this trial, the placebo-treated patients steadily regained weight, maintaining only 20% of their weight loss at the end of the trial. In contrast, the subjects treated with sibutramine maintained their weight for 12 months and then regained an average of only 2 kg, thus maintaining 80% of their initial weight loss after 2 years (James et al., 2000). Despite the higher weight loss with sibutramine at the end of the 18 months of controlled observation, the blood pressure levels of the sibutramine-treated patients were still higher than those of the patients treated with placebo. The final trial for weight maintenance with sibutramine was a study conducted at eight hospitals in the Netherlands in which patients were initially treated before referral to their primary care physicians. A total of 221 patients began the VLCD. Of these patients, 189 lost the required 10 kg during 3 months and were randomized to 10 mg/day of sibutramine or placebo for the remaining 15 months. Mean weight loss during the VLCD period for the successful subjects was 14.5% from baseline. After 2 additional months of treatment in the hospital clinic, the final 13 months were conducted in the general practitioner s offices. At 18 months, the odds ratio was 1.76 (95% CI ) favoring weight loss with sibutramine (P 0.03). With use of the intent-totreat analysis, 80% of the weight loss at the end of the VLCD was maintained by 70, 51, and 30% of those receiving sibutramine at 6, 12, and 18 months compared with 48, 31, and 20% of those receiving placebo, and these differences were significant at all time points (P 0.03) (Mathus-Vliegen, 2005). The possibility of using sibutramine as intermittent therapy has been tested in a randomized, placebo-controlled trial lasting 52 weeks (Wirth and Krause, 2001) (Fig. 7). The patients randomized to sibutramine received one of two regimens. One group received continuous treatment with 15 mg/day for 1 year, and the other group had two 6-week periods when sibutramine was withdrawn. During the periods when the drug was replaced by placebo, there was a small regain in weight that was lost when the drug was resumed. At the end of the trial, the continuous-therapy and intermittent-therapy groups lost the same amount of weight. c. Studies in special populations. i. Diabetic patients. The effects of sibutramine in diabetic patients have been examined in eight studies. In a 3-month (12-week) trial, patients with diabetes who were treated with 15 mg/day of sibutramine lost 2.4 kg (2.8%), compared with a loss of 0.1 kg (0.12%) in the placebo group (Finer et al., 2000). In this study, hemoglobin A 1c levels decreased 0.3% in the drug-treated group and remained stable in the placebo group. Fasting glucose values decreased 0.3 mg/dl in the drug-treated patients and increased 1.4 mg/dl in the placebo-treated group. In a 24-week trial, the dose of sibutramine was increased from 5 to 20 mg/day over 6 weeks (Fujioka et al., 2000). Among those who completed the treatment, weight loss was 4.3 kg (4.3%) in the sibutraminetreated patients, compared with 0.3 kg (0.3%) in placebo-treated patients. Hemoglobin A 1c levels decreased 1.67% in the drug-treated group, compared with 0.53% in the placebo-treated group. These changes in glucose FIG. 6. Effect of sibutramine on weight maintenance. Adapted from James et al. (2000). FIG. 7. Effect on body weight of continuous versus intermittent use of sibutramine in a randomized, placebo-controlled clinical trial. Adapted from The Journal of the American Medical Association, volume 286, pp Copyright 2001 American Medical Association.
12 162 BRAY AND GREENWAY and hemoglobin A 1c levels were expected from the amount of weight loss associated with drug treatment. In a 12-month multicenter, randomized, placebo-controlled study (McNulty et al., 2003), 194 diabetics receiving metformin were assigned to placebo (n 64), 15 mg/day of sibutramine (n 68), or 20 mg/day of sibutramine (n 62). At 12 months, weight loss in the 15 mg/day group was kg and in the 20 mg/day group it was kg compared with kg in the placebo group. Glycemic control improved in parallel with weight loss. Sibutramine raised sitting diastolic blood pressure by 5 mm Hg in 43% of those receiving 15 mg/day of sibutramine compared with 25% for the placebo group (P 0.05). Pulse rate increased 10 beats/min in 42% of those receiving sibutramine, compared with 17% for those receiving placebo. The 2-year trial by Redmon et al. (2005) was a doubleblind, randomized, placebo-controlled trial that had a crossover for the control group after the 1st year. The treatment group received 10 mg/day of sibutramine for the entire 2 years. In addition, they had a portion-controlled diet used for 7 days at the end of each 2 months. There was a significantly greater weight loss in the drug-treated group that reached 9.7 kg at 12 months, compared with 1.6 kg in the placebo-treated group. During the 2nd year, those receiving sibutramine continuously regained weight slowly, rising to a maximal loss of 6.3 kg at 24 months. In contrast, the group that got sibutramine only during the 2nd year weighed less at 24 months than those receiving sibutramine continuously for 2 years ( 6.3 kg in the continuous treatment group versus 9.7 kg for the crossover group). There was an improvement in diabetic control associated with the weight loss. A meta-analysis has been done of eight studies in diabetic patients receiving sibutramine (Vettor et al., 2005). In this meta-analysis, the changes in body weight, waist circumference, glucose, hemoglobin A 1c, triglycerides, and HDL cholesterol favored sibutramine. The mean weight loss was kg for those treated with sibutramine and kg for the placebo-treated patients. There was no significant change in systolic blood pressure, but diastolic blood pressure was significantly higher in the sibutramine-treated patients (Vettor et al., 2005). In the meta-analysis by Norris et al. (2004) the net weight loss over 12 to 26 weeks in 4 trials, including 391 diabetic patients, was 4.5 kg (95% CI kg). ii. Hypertensive patients. Some trials have reported the use of sibutramine to treat overweight patients with hypertension. In a 3-month trial, where all patients received -blockers, with or without thiazides, for their hypertension McMahon et al. (2000) reported that sibutramine-treated patients lost 4.2 kg (4.5%), compared with a loss of 0.3 kg (0.3%) in the placebo-treated group. Mean supine and standing diastolic and systolic blood pressure levels were not significantly different between drug-treated and placebo-treated patients. Heart rate, however, increased by beats/min (mean S.D.) in the sibutramine-treated patients, compared with an increase of beats/min in the placebo group. In another 52-week trial, patients with hypertension whose blood pressure levels were controlled with calcium channel blockers with or without -blockers or thiazides (McMahon et al., 2000) received sibutramine in doses that were increased from 5 to 20 mg/day during the first 6 weeks. Weight loss was significantly greater in the sibutramine-treated patients, averaging 4.4 kg (4.7%), compared with 0.5 kg (0.7%) in the placebotreated group. Diastolic blood pressure levels decreased 1.3 mm Hg in the placebo-treated group and increased 2 mm Hg in the sibutramine-treated group. Systolic blood pressure levels increased 1.5 mm Hg in the placebo-treated group and 2.7 mm Hg in the sibutraminetreated group. Heart rate was unchanged in the placebotreated patients, but increased by 4.9 beats/min in the sibutramine-treated patients. The effects of sibutramine on blood pressure have been evaluated in a meta-analysis of 21 studies by Kim et al. (2003). Sibutramine produced a significant overall weight loss and a significant increase in both systolic and diastolic blood pressures. In a subgroup analysis, they found the effect on systolic blood pressure to be greater with higher doses of sibutramine, in individuals weighing 92 kg and in younger individuals ( 44 years of age). Older individuals with body weights of 92 kg also showed a greater rise in diastolic blood pressure. In another analysis of two studies with use of sibutramine for 48 weeks Jordan et al. (2005) reported that sibutramine significantly reduced body weight but did not lead to a difference in systolic blood pressure after 48 weeks ( mm Hg for placebo versus mm Hg for the sibutramine group). However, the change in diastolic blood pressure was statistically significant with a small rise of mm Hg in the sibutramine group and a decrease of mm Hg in the placebo group (P 0.049). iii. Sibutramine plus behavioral weight loss. Sibutramine has been studied as part of a behavioral weight-loss program in two reports (Wadden et al., 2005). With sibutramine alone, the weight loss over 12 months was kg (5%). Behavior modification alone produced a weight loss of kg. Adding a brief behavioral therapy session to a group that also received sibutramine produced a slightly larger weight loss of kg. When the intensive lifestyle intervention was combined with sibutramine, the weight loss increased to kg. When a structured meal plan was added to the medication and behavioral modification in one of these studies (Wadden et al., 2005), the weight loss increased further to 15 kg (Wadden et al., 2001). Completing the food intake records was a strong predictor of success (Wadden et al., 2005). Those in the combined therapy group receiving an intensive lifestyle program and sibutramine who were in the highest third for record keeping lost kg compared with kg in the lowest third for record-keeping (Fig. 8).
13 DRUG TREATMENT OF OBESITY 163 FIG. 8. Effect on body weight loss of sibutramine alone or with additional supplementary behavior therapies. The study had four groups. Lifestyle was essentially the same as sibutramine brief visit and was left off. Combined therapy was sibutramine lifestyle 30 group visits. Data from Wadden et al. (2005). iv. Studies in children. Sibutramine has also been used in children and adolescents (Berkowitz et al., 2003, 2006; Godoy-Matos et al., 2005). In a single center, 85 adolescents aged 13 to 17 years with a BMI of 32 to 44 kg/m 2 were randomized to treatment for 6 months with placebo or sibutramine. Weight loss in the drug-treated group was 7.8 kg, for an 8.5% reduction in BMI, compared with 3.2 kg in the placebo group, for a 4.0% reduction in BMI. When the placebo group was switched to sibutramine after 6 months, there was an additional significant weight loss in this group. In a 12-month, multicenter, randomized, placebo-controlled trial, 498 adolescents aged 12 to 16 were treated with sibutramine or placebo (Berkowitz et al., 2006). The dose of sibutramine was 10 mg/day for 6 months and then increased to 15 mg/day in those who had not lost 10% of their baseline BMI. After 12 months, the mean absolute change in BMI was 2.9 kg/m 2 ( 8.2%) in the sibutramine group compared with 0.3 kg/m 2 ( 0.8%) in the placebo group (P 0.001). Triglycerides HDL cholesterol, and insulin sensitivity improved, and there was no significant difference in either systolic or diastolic blood pressure. d. Meta-analysis of sibutramine studies. Several meta-analyses of sibutramine have been published (Haddock et al., 2002; Avenell et al., 2004; Li et al., 2005). By pooling four studies Haddock et al. (2002) estimated the weight loss in patients treated with sibutramine as 5.3 kg (range kg) compared with 1.8 kg (range kg) for the placebo-treated groups. In the meta-analysis of Li et al. (2005), the overall mean difference after 12 months of therapy in five studies was 4.45 kg (95% CI 5.29 to 3.62 kg) (Table 7). In the meta-analysis by Avenell et al. (2004), the overall placebo-subtracted effect of sibutramine at 12 months was 4.12 kg (95% CI 4.97 to 3.26 kg). Table 8 shows a summary for each of the trials that had data for sibutramine out to 12 months (Avenell et al., 2004). After an additional interval in the weight maintenance studies, the data showed a loss at 15 months of 3.70 kg (95% CI 5.71 to 1.69 kg) (Apfelbaum et al., 1999) and at 18 months of 3.40 kg (95% CI 4.45 to 2.35 kg) (James et al., 2000). e. Combining sibutramine and orlistat. Because sibutramine works on noradrenergic and serotonergic reuptake mechanisms in the brain and orlistat works peripherally to reduce triglyceride digestion in the GI tract, their mechanisms of action do not overlap and combining them might provide additive weight loss (Wadden et al., 2000). To test this possibility, researchers randomly assigned patients to orlistat or placebo after 1 year of treatment with sibutramine (Wadden et al., 2000). During the additional 4 months of treatment, the two groups lost no significant amount of weight and adding orlistat had no detectable effect. In an open-label randomized 12-week study, 86 overweight patients were assigned to treatment with 120 mg of orlistat three times a day, to treatment with 10 mg/day of sibutramine, the combination of orlistat and sibutramine, or to a diet group. During the 12 weeks, sibutramine produced more weight loss than orlistat alone. In another study, 150 obese subjects were randomized to 850 mg of metformin b.i.d., 120 mg of orlistat t.i.d., or 10 mg of sibutramine b.i.d. treatment for 6 months. The BMI decreased by 9.1, 9.9, and 13.6%, respectively. Weight loss was greater in the sibutramine group than in either the metformin or orlistat groups (P 0.001) (Gokcel et al., 2002). A third study, also of 6 months duration, randomized 89 obese women to orlistat, sibutramine, or the combination and showed weight losses of 5.5, 10.1, and 10.8 kg, respectively. The combination was superior to orlistat but not to sibutramine alone (Sari et al., 2004). A fourth study randomized 86 obese subjects to 10 mg/day of sibutramine, 360 mg/day of orlistat, the combination, or diet alone. In this study, the combination group lost more weight than the diet alone or orlistat groups, but the amount lost was similar to that for sibutramine alone (Aydin et al., 2004). Thus, adding orlistat to sibutramine did not significantly enhance weight loss, confirming the observations of Wadden et al. (2000) and Kaya et al. (2004). TABLE 7 Meta-analysis of net weight loss with sibutramine (placebo-drug) Adapted from Li et al. (2005). Reference Apfelbaum et al. (1999) Smith and Goulder (2001) Hauner et al. (2000) McNulty et al. (2003) Wirth and Krause (2001) Mean 95% CI 5.70 ( 7.77 to 3.63) 3.00 ( 4.55 to 1.45) 5.30 ( 6.83 to 3.77) 4.80 ( 6.02 to 3.58) 4.00 ( 5.01 to 2.99)
14 164 BRAY AND GREENWAY Adapted from Avenell et al. (2004). TABLE 8 Meta-analysis of the effect of sibutramine versus placebo and diet at 12 months Reference Treatment Control n Mean S.D. n Mean S.D. Weight Weighted Mean Difference % 95% CI Apfelbaum et al. (1999) ( 7.77 to 3.63) McMahon et al. (2000) ( 5.75 to 2.05) Smith and Goulder (2001) ( 4.31 to 1.29) Smith and Goulder (2001) ( 6.38 to 3.22) f. Dosage and safety considerations. Sibutramine is available in 5-, 10-, and 15-mg doses; 10 mg per day as a single dose is the recommended starting level, with titration up or down, depending on response. Doses higher than 15 mg/day are not recommended. Of the patients who lost 2 kg (4 lb) in the first 4 weeks of treatment, 60% achieved a weight loss of 5%, compared with 10% of those who did not lose 2 kg (4 lb) in 4 weeks. Combining data from the 11 studies on sibutramine showed a reduction in triglyceride, total cholesterol, and LDL cholesterol levels and an increase in HDL cholesterol levels that were related to the magnitude of the weight loss. 3. Rimonabant. a. Mechanism of action. There are two cannabinoid receptors CB-1 (470 amino acids in length) and CB-2 (360 amino acids in length). The CB-1 receptor has almost all the amino acids that comprise the CB-2 receptor with additional amino acids at both ends. CB-1 receptors are distributed through the brain in the areas related to feeding, on fat cells, in the gastrointestinal tract. The CB-2 receptors are primarily on immune cells. Marijuana and tetrahydrocannabinol, which stimulate the CB-1 receptor, both increase high fat and high sweet food intake. There are two well-characterized endogenous endocannabinoids called anandamide and 2-arachidonylglycerol. When injected into the brain, they also increase food intake. In human beings, fasting increases the levels of these endocannabanoids. The rewarding properties of cannabinoid agonists are mediated through the mesolimbic dopaminergic system. Rimonabant is a specific antagonist of the CB-1 receptor. It inhibits intake of sweet foods by marmosets and reduces intake of high-fat foods in rats. Intake of standard rat chow is not affected by rimonabant. In addition to inhibiting the intake of highly palatable foods, pairfeeding experiments in diet-induced obese rats showed that the rimonabant-treated animals lost 21% of their body weight compared with 14% in the pair-fed controls. This suggests, at least in rodents, that rimonabant increases energy expenditure in addition to reducing food intake. CB-1 knockout mice are lean and resistant to diet-induced weight gain. CB-1 receptors are up-regulated on adipocytes in diet-induced obese mice, and rimonabant increases adiponectin, a fat cell hormone associated with insulin sensitivity (Bensaid et al., 2003; Kirkham, 2005; Juan-Pico et al., 2006; Pagotto et al., 2006). b. Long-term studies. The results of four phase III trials of rimonabant for the treatment of overweight have been published. The first trial, called the Rimonabant in Obesity (RIO)-Europe trial, was reported in 2005 (Van Gaal et al., 2005) and was intended to be conducted in Europe, but slow recruitment led to inclusion of 276 subjects from the United States. This was a 2-year trial with 1-year results reported in this article. A total of 1507 patients with a BMI 30 kg/m 2 without comorbidities or a BMI 27 kg/m 2 with hypertension or dyslipidemia were stratified on whether they lost or 2 kg during the run-in and then were randomized in a ratio of 1:2:2 to receive placebo, 5 mg/day of rimonabant, or 20 mg/day of rimonabant. The energy content of the diet was calculated by subtracting 600 kcal/day from the energy requirements calculated from the Harris-Benedict equation. The trial consisted of a 4-week run-in period followed by 52 weeks of drug treatment. Of those who started, 61% (920) completed the 1st year. Weight loss was 2% in the placebo group and 8.5% in the 20-mg rimonabant group. Baseline weight was between 98.5 kg (placebo group) and kg (for the rimonabant 20-mg/day dose). During the run-in, there was a mean 1.9-kg weight loss. From baseline at the end of the run-in those in placebo the group who completed the trial lost an additional 2.3 kg, the low-dose rimonabant group (5 mg/day) lost 3.6 kg, and the high-dose group (20 mg/day) lost 8.6 kg. On an intent-to-treat basis these numbers were a weight loss of 1.8 kg for the placebo group, 3.4 kg for the 5-mg/day group, and 6.6 kg for the 20-mg/day group. Expressing the data as a responder analysis, the authors reported that 30.5% of the placebo group lost 5%, compared with 44.2% for the 5-mg/day and 67.4% for the 20-mg/day dose of rimonabant. When a weight loss of 10% was considered, the numbers were 12.4% for the placebo group, 15.3% for the 5 mg/day dose of rimonabant group, and 39% for the 20-mg/day dose group. Waist circumference was also reduced by treatment. With the intentto-treat analysis, waist declined 2.4 cm in the placebo group, 3.9 cm in the 5-mg/day dose group, and 6.5 cm
15 DRUG TREATMENT OF OBESITY 165 in the 20-mg/day dose group. Triglycerides were reduced by 6.8% in the 20-mg/day group compared with a rise of 8.3% in the placebo group. HDL cholesterol increased by 22.3% compared with 13.4% in the placebo group. These changes in metabolic parameters were reflected in a change in the prevalence of the metabolic syndrome. Among those in the placebo group who completed the study, there was a 33.9% reduction in the prevalence of the metabolic syndrome, compared with 34.8% in the 5-mg/day rimonabant dose group and 64.8% in the 20-mg/day dose group. In the 20-mg/day group the LDL particle size increased, adiponectin increased, glucose decreased, insulin decreased, C-reactive protein decreased, and the metabolic syndrome prevalence was cut in half. There was no significant change in blood pressure or pulse among the groups. Discontinuation for adverse events was similar, but the reasons were different. Among placebo-treated patients, it was for lack of weight loss. With the higher dose of rimonabant (20 mg/day) depressed mood disorders, nausea, vomiting, diarrhea, headache, dizziness, and anxiety were all more common than in the placebo group. The Hospital Anxiety and Depression scale scores were not significantly different among treatment groups. Slightly more patients withdrew for drug-related adverse events in the 5-mg/day dose group and even more with the 20-mg/day dose relative to those receiving placebo. The major reasons for withdrawal were psychiatric, nervous system, and gastrointestinal track symptoms. The complaints, which occurred with 5% frequency in the drug-treated patients, included upper respiratory tract infection, nasopharyngitis, nausea, influenza, diarrhea, arthralgia, anxiety, insomnia, viral gastroenteritis, dizziness, depressed mood, and fatigue in the 20-mg/day dose group (Pi-Sunyer et al., 2006). The second study was in dyslipidemic patients and was called the Rimonabant in Obesity-Lipids (RIO-Lipids) study (Despres et al., 2005). This was a 12-month randomized, double-blind, placebo-controlled trial of rimonabant at 5- and 20-mg/day doses versus placebo in overweight subjects eating a 600-kcal/day deficit diet. It was conducted at 67 sites in eight countries. As a lipids trial, the inclusion criteria were a BMI of 27 to 40 kg/m 2, elevated fasting triglycerides ( mg/dl), ratio of cholesterol to HDL-cholesterol 5 in men and 4.5 in women, and no more than 5-kg variation in body weight in the previous 3 months. Subjects were stratified at the run-in by triglycerides or 400 mg/dl and at the end of run-in by a weight loss of or 2 kg. Randomization was on a 1:1:1 basis of placebo, 5 mg/day of rimonabant, and 20 mg/day of rimonabant. After the end of the 4-week run-in participants, were randomized and treated for 12 months; the dropout rate was 40% by the end of 12 months. Weight losses in this trial were almost identical to those in the Rio-Europe trial. After an 2-kg weight loss during the run-in, the patients in the placebo group who completed the trial lost an additional 2.3 kg, compared with 4.2 kg in the 5-mg/day rimonabant dose group and 8.8 kg in the 20-mg/day dose group. Waist circumference also showed a dose-dependent reduction of 3.4 cm in the placebo group, 4.9 cm in the 5-mg/day dose group, and 9.1 cm in the 20-mg/day dose group. A number of other metabolic parameters also responded to the drug or weight loss. These included a decrease in triglycerides, an increase in HDL cholesterol, a decrease in peak size of LDL cholesterol particles, an increase in adiponectin, a decline in fasting insulin, a fall in leptin and a decrease in C-reactive protein. Several liver enzymes fell with treatment, suggesting improvement in nonalcoholic steatosis. Blood pressure decreased significantly in RIO-Lipids in contrast with RIO-Europe. As might be expected from these metabolic changes, the prevalence of the metabolic syndrome in those who met the Adult Treatment Panel III criteria at randomization fell to 25.8% in the 20-mg/day group, to 40.0% in the 5-mg/day group, and to 41.0% in the placebo group. The third randomized, double-blind, placebo-controlled study, called RIO-North America, was also a 2-year study in which randomized 3045 overweight subjects with a BMI 30 kg/m 2 or with a BMI 27 kg/m 2 with treated or untreated hypertension or dyslipidemia and without diabetes to placebo, 5 mg/day of rimonabant, or 20 mg/day of rimonabant. Participants were instructed on a 600-kcal/day deficit diet. Randomization and baseline occurred after a 4 week run-in period in which subjects lost an average of 1.9 kg. They were thus stratified by whether they lost or 2 kg during the run-in. At 1 year, half of the patients in each drug group were switched to placebo on the basis of their initial randomization. The trial was conducted at 64 American and 8 Canadian centers. At 1 year, completion rates were 51 to 55% for the three arms. During the 1st year, weight loss was 2.8 kg in the placebo group and 8.6 kg in the 20-mg/day rimonabant group (Fig. 9). Weight loss declined steadily until week 36, after which it plateaued. For the 2nd year, those individuals who were switched from rimonabant to placebo regained weight at almost the mirror image of the rate at which they lost it during the 1st year. At the end of the study they were still slightly lighter, but no different from the group treated with placebo for the full 2 years. Waist circumference decreased, and the percentage with the metabolic syndrome decreased from 34.8 to 21.2% in the 20-mg/day group compared with a change from 31.7 to 29.2% in the placebo group. HDL cholesterol rose more in the rimonabant group treated with 20 mg/day than placebo, and triglycerides fell more in the participants receiving the higher dose of rimonabant. Patients with depression were not included in this study. Adverse events leading to discontinuation of the study were
16 166 BRAY AND GREENWAY higher in the rimonabant-treated than in the placebotreated participants. However, the profile and effectiveness of this agent seem very promising for treatment of obesity and the physical and laboratory findings that make up the metabolic syndrome. c. Studies in special populations. i. Diabetic patients. The fourth study, RIO-Diabetes, was conducted in type 2 diabetic patients. In it, 1045 diabetic subjects treated with diet, metformin, or sulfonylurea drugs at 151 centers in 11 countries were randomized to treatment over 1 year with placebo or rimonabant at 5 or 20 mg/day. Weight loss in the placebo group was 1.4 kg, compared with 2.3 kg in the 5-mg/ day group and 5.3 kg in the 20-mg/day group. Triglycerides and blood pressure declined more in the subjects treated with 20-mg/day rimonabant. Of those who completed the trial, 55.9% lost 5% of body weight during treatment with 20-mg/day rimonabant compared with 19.5% in the placebo-treated group (Scheen et al., 2006). ii. Prevention of weight gain after cessation of smoking. Studies with rimonabant suggest that it can slow or prevent the weight gain of individuals who have quit smoking. d. Safety considerations. Because this drug dampens the feedback systems for pleasurable responses, there is concern about its behavioral effects. To quote from the most recently published article (Scheen et al., 2006), A slightly greater proportion of patients in the rimonabant treatment groups experienced adverse events than did those in the placebo group. The events occurring at a frequency of 5% in the rimonabant group included nausea, diarrhea, vomiting, dizziness, hypoglycemia, fatigue, and anxiety, all of which were generally mild or moderate. Discontinuation for adverse events was also more common in the higher-dose rimonabant group. The most common events for discontinuation were depressed mood disorders, nausea, and dizziness. e. Other cannabinoid antagonists. Other cannabinoid antagonists are said to be under clinical development and may have entered phase III trials. FIG. 9. Effect on body weight of rimonabant in a 2-year, randomized, placebo-controlled clinical trial. Adapted from The Journal of the American Medical Association, volume 295, pp Copyright 2006 American Medical Association. B. Drugs Approved by the U.S. Food and Drug Administration for Short-Term Treatment of Overweight Patients 1. Sympathomimetic Drugs. a. Mechanism of action. Phentermine, diethylpropion, benzphetamine, and phendimetrazine behave in many ways like the adrenergic neurotransmitters and are thus called sympathomimetic amines (Bray and Greenway, 1999). They were originally thought to release norepinephrine from vesicular stores, but more recent data suggest that they act primarily to inhibit reuptake of norepinephrine and dopamine at nerve endings. b. Clinical studies. Most of the data on phentermine, diethylpropion, benzphetamine, and phendimetrazine come from short-term trials (Bray and Greenway, 1999). One of the longest of these clinical trials lasted 36 weeks and compared placebo treatment with continuous phentermine or intermittent phentermine (Munro et al., 1968). Both continuous and intermittent phentermine therapy produced more weight loss than placebo. In the drug-free periods, the patients treated intermittently slowed their weight loss, only to lose weight more rapidly when the drug was reinstituted. In their analysis of pharmacotherapy for treatment of the overweight patient, Haddock et al. (2002) compiled data on both phentermine and diethylpropion. They found that in six studies, phentermine produced a mean weight loss of 6.3 kg (range 3.6 to 8.8 kg) compared with a placebo-induced weight loss of 2.8 kg (range 1.5 to 5.2 kg). For diethylpropion, the mean weight loss in nine studies was 6.5 kg (range 1.9 to 13.1 kg), and for the placebo group it was 3.5 kg (range 0.4 to 10.5 kg). Similar data for benzphetamine are 4.03 kg (range 1.6 to 7.3 kg) and for placebo 0.73 kg (range 1.3 to 2.0 kg). Phentermine and diethylpropion are classified by the U.S. Drug Enforcement Agency as schedule IV drugs; benzphetamine and phendimetrazine are schedule III drugs. This regulatory classification indicates the U.S. Government s belief that they have the potential for abuse, although this potential seems to be very low. Phentermine and diethylpropion are approved for only a few weeks, which usually is interpreted as up to 12 weeks. Weight loss with phentermine and diethylpropion persists for the duration of treatment, suggesting that tolerance does not develop to these drugs. If tolerance were to develop, the drugs would be expected to lose their effectiveness, and patients would require increased amounts of the drug to maintain weight loss. This does not occur.
17 DRUG TREATMENT OF OBESITY 167 IV. Antidepressant and Antiepileptic Drugs That Produce Weight Loss but Are Not Approved by the U.S. Food and Drug Administration for Weight Loss A. Fluoxetine and Sertraline 1. Mechanism of Action. Fluoxetine and sertraline are both selective serotonin reuptake inhibitors that block the transporters that remove serotonin from the neuronal cleft into the presynaptic space for metabolism by monoamine oxidase or storage in granules. They also reduce food intake. In a 2-week placebo-controlled trial, fluoxetine at a dose of 60 mg/day produced a 27% decrease in food intake (Lawton et al., 1995). 2. Clinical Studies. Both fluoxetine and sertraline are approved by the FDA for treatment of depression. In clinical trials with depressed patients lasting 8 to 16 weeks, sertraline gave an average weight loss of 0.45 to 0.91 kg. Fluoxetine at a dose of 60 mg/day (three times the usual dose for treatment of depression) was evaluated in clinical trials for the treatment of overweight patients by Eli Lilly & Co. A meta-analysis of six studies showed a wide range of results with a mean weight loss in one study of 14.5 kg and a weight gain of 0.40 kg in another (Li et al., 2005). In the meta-analysis by Avenell et al. (2004), the weight loss at 12 months was 0.33 kg (95% CI 1.49 to 0.82 kg). Goldstein et al. (1995) reviewed the trials with fluoxetine that included one 36-week trial in type 2 diabetic subjects, a 52-week trial in subjects with uncomplicated overweight, and two 60-week trials in subjects with dyslipidemia, diabetes, or both. A total of 1441 subjects were randomized to fluoxetine (719) or placebo (722); 522 subjects receiving fluoxetine and 504 subjects receiving placebo completed 6 months of treatment. Weight loss in the placebo and fluoxetine groups at 6 months and 1 year were 2.2 and 4.8 kg and 1.8 and 2.4 kg, respectively. The recovery of 50% of the lost weight during the second 6 months of treatment with fluoxetine makes it inappropriate for the long-term treatment of overweight, which requires chronic treatment. Fluoxetine and sertraline, although not good antioverweight drugs, may be preferred in the depressed obese patient over some of the tricyclic antidepressants that are associated with significant weight gain (Fig. 10). B. Bupropion 1. Mechanism of Action. Bupropion is a norepinephrine and dopamine reuptake inhibitor that is approved for the treatment of depression and for smoking cessation. 2. Clinical Studies for Weight Loss. One clinical use for bupropion has been to prevent weight gain after cessation of smoking. It was thus a potential drug for treatment of overweight patients. In one clinical trial, 50 overweight subjects were randomized to bupropion or FIG. 10. Effect on body weight of fluoxetine from pooled data of randomized, placebo-controlled clinical trials. Redrawn from Goldstein et al. (1995). placebo for 8 weeks with a blinded extension for responders to 24 weeks. The dose of bupropion was increased to a maximum of 200 mg twice daily in conjunction with a calorie-restricted diet. At 8 weeks, 18 subjects in the bupropion group lost % of body weight compared with % for the 13 subjects in the placebo group (P ). After 24 weeks, the 14 responders to bupropion lost % of initial body weight, of which 75% was fat as determined by dual energy X-ray absorptiometry (Gadde et al., 2001). Two multicenter clinical trials, one in obese subjects with depressive symptoms and one in uncomplicated overweight patients, followed this study. In the study of overweight patients with depressive symptom ratings of 10 to 30 on a Beck Depression Inventory, 213 patients were randomized to 400 mg of bupropion per day and 209 subjects to placebo for 24 weeks. The 121 subjects in the bupropion group who completed the trial lost % of initial body weight compared with % in the 108 subjects in the placebo group (P ) (Jain et al., 2002). The study in uncomplicated overweight subjects randomized 327 subjects to 300 mg/day of bupropion, 400 mg/day of bupropion, or placebo in equal proportions. At 24 weeks, 69% of those randomized remained in the study and the percent losses of initial body weight were 5 1, 7.2 1, and % for the placebo, 300 mg/day of bupropion, and 400 mg/day of bupropion groups, respectively (P ). The placebo group was randomized to the 300- or 400- mg/day group at 24 weeks, and the trial was extended to week 48. By the end of the trial, the dropout rate was 41%, and the weight loss in the 300-mg/day bupropion and 400-mg/day bupropion groups was % and % of initial body weight, respectively (Anderson et al., 2002). Thus, it seems that nondepressed subjects may respond to bupropion with weight loss to a greater extent than those with depressive symptoms (Fig. 11). Because bupropion is at the end of its
18 168 BRAY AND GREENWAY FIG. 11. Effect on body weight of bupropion in a randomized, placebocontrolled clinical trial. Adapted from Anderson et al. (2002). patent life, it is not being developed for obesity, but similar drugs in this class would be obvious candidates. C. Topiramate 1. Mechanism of Action. Topiramate is a neurotherapeutic agent that is approved for treatment of selected seizure disorders. It is a weak carbonic anhydrase inhibitor, exhibiting selectivity for carbonic anhydrase isoforms II and IV. Topiramate also modulates the effects at receptors for the GABA-A receptor and the -amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid/ kainate subtype of the glutamate receptor. This drug also exhibits state-dependent blockade of voltage-dependent Na or Ca 2 channels. These mechanisms are believed to contribute to its antiepileptic properties. The modulation of GABA receptors may provide one potential mechanism to reduce food intake, although other mechanisms, yet to be described, may be more important in defining its effects on body weight (Astrup and Toubro, 2004). The final words on the mechanism(s) by which topiramate reduces food intake and body weight are yet to be published. 2. Clinical Studies for Weight Loss. Topiramate is an antiepileptic drug that was discovered to give weight loss in the clinical trials for epilepsy. Weight losses of 3.9% of initial weight were seen at 3 months and losses of 7.3% of initial weight were seen at 1 year (Ben- Menachem et al., 2003). Bray et al. (2003) reported a 6-month, placebo-controlled, dose-ranging study of 385 obese subjects who were randomized to placebo or topiramate at 64, 96, 192, or 384 mg/day. These doses were gradually reached by a tapering increase and were reduced in a similar manner at the end of the trial. Weight loss from baseline to 24 weeks was 2.6, 5, 4.8, 6.3, and 6.3% in the placebo, 64-, 96-, 192-, and 384-mg/day groups, respectively. The most frequent adverse events were paresthesias, somnolence and difficulty with concentration, memory, and attention. This trial was followed by two multicenter trials. The first trial randomized 1289 obese subjects to topiramate 89, 192, or 256 mg/day. This trial was terminated early because of the sponsor s decision to pursue a time-release form of the drug. The 854 subjects who completed 1 year of the trial before it was terminated lost 1.7, 7, 9.1, and 9.7% of their initial body weight in the placebo, 89-, 192-, and 256-mg/day groups, respectively. Subjects in the topiramate groups had significant improvement in blood pressure and glucose tolerance (Wilding et al., 2004). The second trial enrolled 701 subjects who were treated with a very low-calorie diet to induce an 8% loss of initial body weight. The 560 subjects who achieved an 8% weight loss were randomized to 96 or 192 mg/day of topiramate or placebo. The sponsor also terminated this study early. At the time of termination, 293 subjects had completed 44 weeks. The topiramate groups lost 15.4 and 16.5% of their baseline weight whereas the placebo group lost 8.9% (Astrup et al., 2004). Although topiramate is still available as an antiepileptic drug, the development program to obtain an indication for overweight was terminated by the sponsor because of the associated adverse events (Fig. 12). 3. Special Situations. Topiramate has also been evaluated in the treatment of binge-eating disorder. Thirteen women with binge-eating disorder were treated in an open-label study using a mean dose of 492 mg/day of topiramate. The binge-eating disorder symptoms improved, and weight loss was observed (Shapira et al., 2000). This open-label study was followed by a randomized, controlled trial of 14 weeks in subjects with binge-eating disorder. Sixty-one subjects were randomized to 25 to 600 mg/day of topiramate or placebo in a 1:1 ratio. The topiramate group had improvement in binge eating symptoms and lost 5.9 kg at an average topiramate dose of 212 mg/day (McElroy et al., 2003). The 35 subjects who completed this trial were given the opportunity to participate in an open-label extension. The topiramate-treated subjects continued to maintain improvement in binge-eating symptoms and weight (McElroy et al., 2004b). FIG. 12. Effect on body weight of topiramate in a double-blind, randomized, placebo-controlled dose-ranging clinical trial. Adapted from Bray et al. (2003).
19 DRUG TREATMENT OF OBESITY 169 Topiramate has also been used to treat patients with the Prader-Willi syndrome. Three subjects with Prader- Willi syndrome were treated with topiramate and had a reduction in the self-injurious behavior that is associated with this uncommon genetic disease (Shapira et al., 2002). A second study in seven additional subjects confirmed these findings (Smathers et al., 2003). A third study evaluated appetite, food intake, and weight. Although the self-injurious behavior improved, there was no effect on these other parameters (Shapira et al., 2004). Topiramate was also used to treat two subjects with nocturnal eating syndrome and two subjects with sleep-related eating disorder. There was an improvement in all subjects and a 11-kg weight loss over 8.5 months with an average topiramate dose of 218 mg/day (Winkelman, 2003). D. Zonisamide 1. Mechanism of Action. Zonisamide is an antiepileptic drug that has serotonergic and dopaminergic activity in addition to inhibiting sodium and calcium channels. The mechanism by which it lowers body weight has not been clarified. 2. Clinical Studies for Weight Loss. Weight loss was noted in the clinical trials for the treatment of epilepsy, again suggesting a potential agent for weight loss. Gadde et al. (2003) tested this possibility by performing a 16-week randomized, controlled trial in 60 obese subjects. Subjects were placed on a calorie-restricted diet and randomized to zonisamide or placebo (Fig. 13). The zonisamide was started at 100 mg/day and increased to 400 mg/day. At 12 weeks, the dose of zonisamide for those subjects who had not lost 5% of initial body weight was increased to 600 mg/day. The zonisamide group lost 6.6% of initial body weight at 16 weeks compared with 1% in the placebo group. Thirty-seven subjects completing the 16-week trial elected to continue to week 32. Of these, 20 were in the zonisamide group and 17 were in the placebo group. At the end of 32 weeks, the 19 FIG. 13. Effect on body weight of zonisamide in a randomized, placebo-controlled clinical trial. Redrawn from The Journal of the American Medical Association, volume 289, pp Copyright 2003 American Medical Association. subjects who remained in the zonisamide group lost 9.6% of their initial body weight compared with 1.6% for the 17 subjects in the placebo group. McElroy et al. (2004a) evaluated zonisamide in an open-label prospective trial in subjects with binge-eating disorder. Fifteen subjects were treated with doses of 100 to 600 mg/day for 12 weeks. The eight subjects who completed the trial had an average dose of 513 mg/day, experienced an improvement in their binge-eating symptoms and lost a significant amount of weight. E. Lamotrigine 1. Mechanism of Action. Lamotrigine is an antiepileptic drug that does not produce weight gain (Devinsky et al., 2000). 2. Clinical Studies. A recent double-blind, randomized, placebo-controlled trial examined the effects of placebo versus lamotrigine escalated from 25 to 200 mg/day on weight loss in a 26-week trial with 40 healthy overweight (BMI kg/m 2 ) adults 18 years of age. At the end of the trial body weight was marginally lower (P 0.062) in the lamotrigine-treated group ( 6.4 kg) than in the placebo-treated group ( 1.2 kg) (Merideth, 2006). V. Drugs Approved by the U.S. Food and Drug Administration for Uses Other Than Overweight A. Metformin 1. Mechanism of Action. Metformin is a biguanide that is approved for the treatment of diabetes mellitus, a disease that is exacerbated by overweight and weight gain. This drug reduces hepatic glucose production, decreases intestinal absorption from the gastrointestinal tract, and enhances insulin sensitivity. 2. Clinical Studies. In clinical trials in which metformin was compared with sulfonylureas, it produced weight loss (Bray and Greenway, 1999). In one French trial, Biguanides and Prevention of Risks in Obesity (BIGPRO), metformin was compared with placebo in a 1-year multicenter study in 324 middle-aged subjects with upper body adiposity and the insulin resistance syndrome (metabolic syndrome). The subjects receiving metformin lost significantly more weight (1 2 kg) than the placebo group, and the conclusion of the study was that metformin may have a role in the primary prevention of type 2 diabetes (Fontbonne et al., 1996). In a meta-analysis of three of these studies Avenell et al. (2004) reported a weighted mean weight loss at 12 months of 1.09 kg (95% CI 2.29 to 0.11 kg). The best trial of metformin is the Diabetes Prevention Program study of individuals with impaired glucose tolerance. The main part of this study included three treatment arms to which participants were randomly assigned, if they were 25 years of age, had a BMI 24 kg/m 2 (except Asian-Americans who only needed a BMI 22 kg/m 2 ), and had impaired glucose tolerance. The three primary arms included lifestyle (n 1079 partic-
20 170 BRAY AND GREENWAY ipants), metformin (n 1073), and placebo (n 1082). During the follow-up averaging 3.2 years, the metformin-treated group lost 2.5% of their body weight (P 0.001, compared with placebo), and the conversion from impaired glucose tolerance to diabetes was reduced by 30% compared with placebo (Fig. 14). In the DPP trial, metformin was more effective in reducing the development of diabetes in the subgroup who were most overweight and in the younger members of the cohort (Knowler et al., 2002). Although metformin does not produce enough weight loss (5%) to qualify as a weight-loss drug according to the FDA criteria, it would seem to be a very useful choice for overweight individuals who have diabetes or are at high risk for diabetes. One area for use of metformin is treatment of women with the polycystic ovary syndrome in whom a modest weight loss may contribute to increased fertility and reduced insulin resistance (Ortega- Gonzalez et al., 2005). Metformin (850 mg) b.i.d. and 120 mg of orlistat t.i.d. were compared with 10 mg of sibutramine daily in obese subjects over 6 months. BMI decreased by 9.1, 9.9, and 13.6%, respectively. Weight loss was greater in the sibutramine group than in either the metformin or orlistat groups (P 0.001) (Gokcel et al., 2002). B. Pramlintide 1. Mechanism of Action. Amylin is peptide found in the -cell of the pancreas. It is secreted along with insulin, and circulates in the blood. Amylin is deficient in type I diabetes in which -cells are immunologically destroyed. Pramlintide is a synthetic amylin analog with a prolonged biological half-life (Riddle and Drucker, 2006). 2. Clinical Studies. Pramlintide is approved by the FDA for the treatment of diabetes. Unlike insulin, sulfonylureas, and thiazolidinediones, pramlintide is associated with weight loss. In a study of 651 subjects with type 1 diabetes randomized to placebo or subcutaneous pramlintide 60 g three or four times a day along with an insulin injection, the hemoglobin A 1c decreased FIG. 14. Effect on body weight of metformin in a double-blind, placebo-controlled, randomized trial from the Diabetes Prevention Program. Data from Knowler et al. (2002) to 0.34% and weight decreased 1.2 kg in the pramlintide group relative to placebo (Ratner et al., 2004). Maggs et al. (2003) analyzed the data from two 1-year studies in insulin-treated type II diabetic subjects randomized to pramlintide 120 g twice a day or 150 g three times a day. Weight decreased by 2.6 kg and hemoglobin A 1c decreased 0.5%. When weight loss was then analyzed by ethnic group, African Americans lost 4 kg, Caucasians lost 2.4 kg, and Hispanics lost 2.3 kg, and the improvement in diabetes correlated with the weight loss, suggesting that pramlintide is effective in ethnic groups with the greatest burden from overweight. The most common adverse event was nausea, which was usually mild and confined to the first 4 weeks of therapy. C. Exenatide 1. Mechanism of Action. GLP-1 is derived from the processing of the preproglucagon peptide, which is secreted by L-cells in the terminal ileum in response to a meal. Increased GLP-1 inhibits glucagon secretion, stimulates insulin secretion, stimulates gluconeogenesis, and delays gastric emptying (Patriti et al., 2004). It has been postulated to be responsible for the superior weight loss and superior improvement in diabetes seen after gastric bypass surgery for overweight (Greenway et al., 2002; Small and Bloom, 2004). GLP-1 is rapidly degraded by dipeptidyl peptidase-4, an enzyme that is elevated in the obese. Bypass operations for overweight increase levels of GLP-1, but do not change the levels of dipeptidyl pepidase-4 (Lugari et al., 2004; Riddle and Drucker, 2006). Exenatide (exendin-4) is a 39-amino acid peptide that is produced in the salivary gland of the Gila monster lizard. It has 53% homology with GLP-1, but it has a much longer half-life. Exenatide decreases food intake and body weight gain in Zucker rats while lowering hemoglobin A 1c (Szayna et al., 2000). It also increases -cell mass to a greater extent than would be expected for the degree of insulin resistance (Gedulin et al., 2005). Exendin-4 induces satiety and weight loss in Zucker rats with peripheral administration and crosses the bloodbrain barrier to act in the central nervous system (Rodriquez de Fonseca et al., 2000; Kastin and Akerstrom, 2003). 2. Clinical Studies with Weight Loss as a Component. Exenatide has been approved by the U.S. Food and Drug Administration for treatment of type 2 diabetic patients whose diabetes is inadequately controlled with either metformin or sulfonylureas. In humans, exenatide reduces fasting and postprandial glucose levels, slows gastric emptying, and decreases food intake by 19% (Edwards et al., 2001). The side effects of exenatide in humans are headache, nausea, and vomiting that are lessened by gradual dose escalation (Fineman et al., 2004). Several clinical trials of 30 weeks duration using exenatide at 10 g s.c./day or a placebo have been reported (Buse et al., 2004; DeFronzo et al., 2005; Kendall